Gesture-Based Interface for Manipulator-based Mobile Robot Control

Abstract

The focus of this paper is on a gesture-based interface designed for controlling a mobile robot equipped with a manipulator. This interface employs a camera to track a person's movements and recognize gestures that involve arm motion. To enable the robot to follow a person reliably through environments with changing lighting conditions, a fast, adaptive tracking algorithm is utilized. The paper compares two alternative approaches for gesture recognition: a template-based approach and a neural network approach. Both methods are combined with the Viterbi algorithm to recognize arm motion-based gestures in addition to static arm poses. The results of this study are presented in the context of an interactive clean-up task, where a person directs the robot to specific locations that require cleaning and instructs it to pick up trash. Service robotics is currently an important area of research, with significant potential for societal impact. As service robots interact directly with people, developing intuitive and user-friendly interfaces is crucial. While prior work has primarily focused on navigation and manipulation, few robotic systems possess flexible user interfaces that enable natural and easy-to-use control of the robot.